Skip to main content
. 2000 Jun 24;320(7251):1693–1696. doi: 10.1136/bmj.320.7251.1693

Table 3.

Logistic regression model for prediction of resolved disease

Variable Level β P value Odds ratio (95% CI)
Intercept 0.68 0.128
Smear test result on first visit No dyskaryosis v dyskaryosis −1.22 <0.001 0.30 (0.20 to 0.43)
Colposcopic opinion Normal v abnormal −0.58 0.007 0.56 (0.36 to 0.85)
Age Continuous 0.03 0.001 1.03 (1.01 to 1.05)
No of pregnancies Continuous −0.12 0.020 0.89 (0.80 to 0.98)
Smoking history Never smoker v ever smoker −0.41 0.035 0.67 (0.46 to 0.97)

Model accuracy: 68% correct; sensitivity 70%; specificity 65%; false positive rate 30%; false negative rate 36%.